COHERENT-LESS DSP

Field of the Invention

The present invention relates to the field of optical communication systems. More particularly, the invention relates to a method and system for receiving and processing high data rate optical communication signals using components of non -coherent technology.

Background of the Invention

The constant growth in the demand for high bandwidth data transmission leads to larger challenges that should be resolved in the physical layer, and particularly by optical transmission technology.

The current high end transmission data rates are in the range of hundreds of G bits/sec. One emerging technology that can support such bitrates for long distances (hundreds of kilometers and above) is coherent transmission and detection. However, coherent technology is perceived as an expensive solution that does not fit the pricing levels that are suitable for sub- hundred kilometers applications of metro-edge in metropolitan area networks and data centers interconnections.

On the other hand, direct detection optical technology (which is non ^¬ coherent technology) is a lower cost and simpler alternative. However, this alternative is limited to lower bit rates and/or shorter distances, and therefore, offers less spectra utilization. Practically, transmission in data rates of lOOGbit/sec is limited to 10km-40km, and requires 4 wavelengths that consume 200GHz of spectral bandwidth.

Yet, there is a very high demand to support 100G transmissions to distances of 100km, either for upgrading existing 10G metro networks, or to support the new demand for data centers interconnections. In addition, better spectrum utilization is required in many of such applications.

"Electrical domain compensation of optical dispersion" (J. McNicol et al, Optical Fiber Communication Conference, Technical Digest OFC/NFOEC, Vol. 4, March 2005) proposes the possibihty of pre-compensation. However, this solution requires using complicated three-arms optical E-field modulator with separate Γ and Q- drives.

It is therefore an object of the present invention to provide a method and system for supporting data transmission in rates of hundreds of Gbit/sec to distances in the order of 100km, based on low cost direct (non-coherent) detection components.

It is another object of the present invention to provide a method and system for compensating chromatic dispersion distortions in optical channels, while maintains the advantage of coherent technology.

Other objects and advantages of the invention will become apparent as the description proceeds.

Summary of the Invention

The present invention is directed to a method of pre-compensating distortions of a time varying optical communication channel. At the transmitter side, the method comprises the following steps :

a) carrying out data multiplexing of the input data streams by a multiplexer/de -multiplexer to convert the rate of the input data streams to another desired rate at a plurality of data lanes with resulting symbols!

b) encoding the symbols in the lanes, using a sampling rate being equal to or greater than the symbol rate; c) performing fixed pre-compensation of deterministic distortions (that may be related to CD) in the channel by pre-filtering the incoming signal to reduce the aliasing effect and convolving the outcome in time or in frequency domain, with the inverse transfer function of the channel, to obtain a CD pre-compensated complex symbol; and

d) performing adaptive pre-compensation of the channel, based on feedback from the channel, side information or a combination thereof.

At the receiver side, distortions post-compensation comprises the following steps:

a) adjusting the amplitude of the received signal to be at full swing in each data lane;

b) digitizing the amplitude of the received signal to obtain a digital signal;

c) performing adaptive ISI compensation of the symbols in each data lane;

d) decoding the compensated symbols in the lanes to obtain a multiplexed data stream; and

e) de-multiplexing the multiplexed data stream to reconstruct the transmitted input data streams.

f) performing fixed pre-compensation of deterministic distortions in the channel by pre-filtering the incoming signal to reduce the aliasing effect; and

g) performing adaptive post-compensation of the channel, based on blind channel estimation, channel feedback, side information or any combination thereof. Adaptive post-compensation may be performed by using MLSE.

In one aspect, Forward Error Correction (FEC) is performed for controlling errors in the data transmission over the channel.

In one aspect, phase pre-compensation may be implemented using a single dual-drive Mach-Zehnder Modulator (MZM), the MZM is not operating in push-pull configuration.

In one aspect, combined signal and phase compensation signal may be generated and used as a driving signal into the MZM.

An arbitrary complex signal with arbitrary amplitude and phase may be generated by a single dual-drive MZM, for CD phase pre-compensation.

High order PAM may be used for transmission, to increase the transmission distance or reduce the number of transmission lanes.

In one aspect, digitation is performed using a sampling rate of 1 sample per symbol. Also, high order PAM may be used in combination with using 1 sample per symbol ADC and consequent baud rate DSP at the receiver side.

PAM transmission levels may be optimized to minimize Amplified Stimulated Emission (ASE) effects.

The present invention is also directed to a system for pre-compensating distortions of the transmitter side of a time varying optical communication channel, which comprises : a) a multiplexer/de -multiplexer for data multiplexing input data streams to convert the rate of the input data streams to another desired rate at a plurality of data lanes with resulting symbols;

b) a processing circuitry for encoding the symbols in the lanes, using a sampling rate being equal to or greater than the symbol rate;

c) a processing circuitry for performing fixed pre-compensation of deterministic distortions in the channel by pre-filtering the incoming signal to reduce the aliasing effect and convolving the outcome in time or in frequency domain, with the inverse transfer function of the channel, to obtain a CD pre-compensated complex symbol; and

d) a processing circuitry for performing adaptive pre-compensation of the channel, based on channel feedback, side information or a combination thereof.

The present invention is further directed to a system for post- compensating distortions at the receiver side of time varying optical communication channel, which comprises :

a) a processing circuitry for adjusting the amplitude of the received signal to be at full swing in each data lane;

b) a sampling circuitry for digitizing the amplitude of the received signal to obtain a digital signal;

c) a processing circuitry for performing adaptive ISI compensation of the symbols in each data lane;

d) a decoder for decoding the compensated symbols in the lanes to obtain a multiplexed data stream; and

e) a de -multiplexer de -multiplexing the multiplexed data stream to reconstruct the transmitted input data streams. f) a processing circuitry for performing fixed pre-compensation of deterministic distortions in the channel by pre-filtering the incoming signal to reduce the aliasing effect and

g) a processing circuitry for performing adaptive post-compensation of the channel, based on blind channel estimation, channel feedback, side information or any combination thereof.

Brief Description of the Drawings

The above and other characteristics and advantages of the invention will be better understood through the following illustrative and non -limitative detailed description of preferred embodiments thereof, with reference to the appended drawings, wherein:

Fig. la shows a principal ASIC block diagram for 2x50Gbit/sec PAM-4 transmission, using the method proposed by the present invention;

Fig. lb shows a principal ASIC block diagram for lx33.25Gbit/sec PAM-8 transmission, using the method proposed by the present invention;

Fig. 2a shows the transmitter's E/O frontend unit for PAM-4, using the method proposed by the present invention;

Fig. 2b shows the transmitter's E/O frontend unit for PAM-8, using the method proposed by the present invention;

Fig. 3a shows the receiver's O/E fronted unit for PAM-4;

Fig. 3b shows the receiver's O/E fronted for PAM-8;

Fig. 4a shows an eye diagram, obtained by the proposed method, using 56Gbit/sec-PAM"4 (with speed-up) modulation, after 80km transmission in amplified link at OSNR value of 35dB;

Fig. 4b shows an eye diagram, obtained by the proposed method, using 112Gbit/sec-PAM-8 (with speed-up) modulation, after

80km transmission in amplified link at OSNR value of 35dB; and Fig. 5 shows the BER performance of the proposed method for twice oversampled PAM-4 as a function of the optical input power.

Detailed Description of Preferred Embodiments

The present invention discloses a method and system for receiving and processing high data rate optical communication signals using components of non-coherent technology. The proposed method allows supporting transmission of hundreds of Gbit/sec to distances in the order of 100km, based on low cost direct detection (non-coherent) components. The proposed method maintains the advantage of coherent technology in the sense of compensation for the chromatic dispersion distortions, by means of phase pre-compensation.

Here the proposed system is based on a compensation approach which uses simple low-cost commercial dual-drive Mach Zehnder modulator (MZM). This compensation approach allows further cost reduction and better spectrum utilization, by increasing the order of modulation from binary Non Return to Zero (NRZ - which is a form of digital data transmission in which the binary low and high states, represented by numerals 0 and 1, are transmitted by specific and constant DC voltages) to high order Pulse Amplitude Modulation (PAM). As a result, the solution proposed by the present invention allows supporting higher spectral efficiency (similar to coherent technology), reducing the number of optical transmitters and receivers elements (thus reducing cost), and yet employing low-cost direct detection components, which are used in conventional direct detection.

According to the proposed method, Digital Signal Processing (DSP) is used at the transmitter (Tx) side, in order to carry out the pre-compensation of the Chromatic Dispersion (CD). The CD pre-compensation phase can be done either in time or frequency domains, at the symbol rate or at a higher rate. When the pre-compensation is carried out at the symbol rate, Anti- Aliasing Filter (AAF) is required. This AAF introduces Inter Symbol Interference (ISI), which is compensated at the receiver (Rx) side, for example by means of the Maximum Likelihood Sequence Estimation (MLSE) equalizer (any other ISI mitigation techniques are also possible).

The pre-compensation is executed by generating an analog electrical signal that drives a Mach-Zehnder Modulator (MZM). The MZM output generates a complex optical signal that consists of a combination of the information data, along with the CD phase pre-compensation.

The proposed pre-compensation can be used for with advanced modulation formats, such as PAM-4 (modulation using two-bit), or PAM-8 (modulation using three-bit). This allows reducing the number of Opto -Electronic (O-E) components, due to the ability to transmit several bits per symbol, which, in turn, leads to significant cost reduction.

Fig. la shows a principal ASIC block diagram for 2x50Gbit/sec PAM-4 transmission, using the method proposed by the present invention. In Fig. la, the transmitter (Tx) section 10a consists of a 10^4 gearbox 101a (a gearbox is essentially a kind of multiplexer/demultiplexer that is used to convert multiple serial data streams at one rate to multiple streams at another rate), 2 PAM-4 mappers 102a, followed by CD pre-compensation blocks 103a, MZM predistortion (it is a specific compensation for the MZM modulator distortion due to its non-linearity) blocks 104a (for imperfect MZM transfer function), 4 Digital to Analog Converters (DACs) 105a and 4 AAFs (not shown). The 10^4 gearbox 101a carries out the bit redistribution (or data multiplexing), for example it takes 10 streams of lOGbit/sec each and produces 4 streams of 25Gbit/sec each. The subsequent PAM-4 mappers perform the PAM-4 encoding on each pair of lanes, with an option to increase the sampling rate of the resulting symbols. The CD compensation block 103a, in turn, carries out two major functions: optional digital pre-filtering of the incoming signal to reduce the aliasing effect (for 1 sample per symbol processing) and convolving the outcome, either in time or in frequency domain, with the inverse transfer function of the channel, which represents the Chromatic Dispersion (CD) phenomenon. The outcome of each CD compensation block 103 contains two tributaries, which are the components of the CD pre-compensated complex symbol. The two tributaries are fed into the MZM preprocessing blocks 104a that generate the digital driving signals of the MZM. In turn, these signals are converted to the analog domain by DACs 105a, which can be used (up to amplification) as a driving signal for the dual-drive (not push pull) MZMs.

The receiver (Rx) section 11a consists of the pair of input variable gain amplifiers 106a, which adjust the received signal into the Analog to Digital Converter (ADC) 107a at full swing. At the next step, the two lanes are digitized, and fed to the MLSE blocks 108a for further ISI compensation. The main MLSEs task is to mitigate the ISI coming from the preceding equivalent channel. The impairments of this equivalent channel include PMD, bandwidth limitations of the Electrical-to-Optical conversion (E/O) and OpticaHo-Electrical conversion (O/E) components, low-pass components such as drivers, modulators, photodiodes, VGAs, etc., as well as any other channel distortions of the actual real channel, except for CD.

Special attention should be paid to the case of CD and MZM compensation that is done at the symbol rate. In any case, the MLSE is adaptive and performing adaptive compensation of said channel, based on feedback from the channel or on side information (i.e., control information sent by the transmitter that is not part of the payload data), automatically tunes itself to the ISI that is present in the equivalent channel. After the MLSE compensation is performed, the symbols are de-mapped by the 2 PAM-4 de- mappers 109a back to bits, which are redistributed by the 4:10 gearbox 110a to the desired number of lanes (which is 10 in this case).

Fig. lb shows a principal ASIC block diagram for lx33.25Gbit/sec PAM-8 transmission, using the method proposed by the present invention. Fig. lb shows a principal ASIC block diagram for 3x33.25Gbit/sec PAM-8 transmission, using the method proposed by the present invention. In Fig. lb, the transmitter (Tx) section 10b consists of a 10:3 gearbox 101b, 2 PAM-8 mappers 102a, followed by CD pre-compensation blocks 103b, a MZM pre -distortion blocks 104b (for imperfect MZM transfer function), 2 Digital to Analog Converters (DACs) 105b and 2 AAFs (not shown). The 10:3 gearbox 101b carries out the bit redistribution (or data multiplexing), for example it takes 10 streams of lOGbit/sec each and produces 3 streams of 33.25Gbit/sec each. The subsequent PAM-8 mappers perform the PAM-8 encoding on each pair of lanes, with an option to increase the sampling rate of the resulting symbols. The CD compensation block 103b, in turn, carries out two major functions: optional digital pre-filtering of the incoming signal to reduce the aliasing effect (for 1 sample per symbol processing) and convolving the outcome, either in time or in frequency domain, with the inverse transfer function of the channel, which represents the Chromatic Dispersion (CD) phenomenon. The outcome of each CD compensation block 103 contains two tributaries, which are the components of the CD pre-compensated complex symbol. The two tributaries are fed into the MZM preprocessing blocks 104b that generate the digital driving signals of the MZM. In turn, these signals are converted to the analog domain by DACs 105b, which can be used (up to amplification) as a driving signal for the dual-drive (not push pull) MZMs.

The receiver (Rx) section lib consists of an input variable gain amplifier 106b, which adjusts the received signal into the Analog to Digital Converter (ADC) 107b at full swing. At the next step, the lane is digitized, and fed to the MLSE block 108b for further ISI compensation.

After the MLSE compensation is performed, the symbols are de-mapped by the PAM-8 de-mapper 109b back to bits, which are redistributed by the 3^10 gearbox 110b to the desired number of lanes (which is 10 in this case).

The ASIC may also include a forward error correction (FEC) functionality (not shown in the figure).

Fig. 2a shows the transmitter's E/O frontend 20a unit for PAM-4, used in the method described above and located immediately after the DAC 105a that is shown in Fig. la. It consists of four drivers 201a, two dual-drive (not push-pull) MZMs 202a, two CW lasers 203a operating at different wavelengths and an Arrayed Waveguide Grating (AWG) multiplexer 204a (or any other optical multiplexer) which combines the resulting optical signals.

Fig. 2b shows the transmitter's E/O frontend 20b unit for PAM-8 or any other single lane solution, used in the method described above and located immediately after the DAC 105b that is shown in Fig. lb. It consists of two drivers 201b, a single-drive (not push-pull) MZM 202b, and a CW laser 203b operating at different wavelengths. The drivers in the E/O Tx front end may be omitted if the DAC output voltage is properly designed. The modulating signals into the MZM are independent signals (not in push pull configuration).

The O/E receiver's (Rx) fronted 30 for PAM-4 is shown in Fig. 3a. It is located at the input of the VGA 106a and ADC 107a that are shown in Fig. la and consists of an AWG demultiplexer 301, followed by two photo- diodes 302a, with two trans-impedance amplifiers 303a at their output. The O/E receiver's (Rx) fronted 31 for PAM-8 or any other single lane solution is shown in Fig. 3b. It is located at the input of the VGA 106b and ADC 107b that are shown in Fig. lb and consists of a single photo-diode 302b with a trans-impedance amplifier 303b.

Fig. 4a shows an eye diagram, obtained by the proposed method, using 56Gbit/sec-PAM"4 (with speed-up) modulation, after 80km transmission in amplified link at OSNR value of 35dB.

Fig. 4b shows an eye diagram, obtained by the proposed method, using 112Gbit/sec-PAM-8 (with speed-up) modulation, after 80km transmission in amplified link at OSNR value of 35dB. Both preliminary results were simulated with the following simulation parameters:

Modulation format: NRZ (PAM2/OOK)

Bit rate: 4x25Gbit/sec=100Gbit/sec

Driver: 4th order Butterworth with 28GHz cutoff frequency

MZM: 2nd order Bessel with 25GHz cutoff frequency

Rx: 4th order Bessel filter with 20GHz cutoff frequency

VGA filter: 2nd order Butterworth with 25GHz cutoff frequency

DAC AAF filter: 4th order Bessel filter with 25GHz cutoff frequency

It can be seen that the eye is open even in the case of 80km transmission, due to implementing the CD pre-compensation method described above, which is done by using twice the symbol rate oversampled signal at the transmitter (Tx). Since all the E/O and O/E components have enough BW (according to the parameters described above), the resulting ISI is relatively small. Therefore, in this case, a simple hard decision slicer with optimized threshold will provide satisfactory results. Fig. 5 shows the BER performance of the proposed method for twice oversampled PAM-4 as a function of the optical input power. It is clearly seen that the proposed method results in negligible penalty in the case of 80 km transmission, compared to back to back transmission (0 km) in the above scenario.

The above examples and description have of course been provided only for the purpose of illustration, and are not intended to limit the invention in any way. As will be appreciated by the skilled person, the invention can be carried out in a great variety of ways, employing more than one technique from those described above, other than used in the description, all without exceeding the scope of the invention.